A data-independent acquisition (DIA)-based quantification workflow for proteome analysis of 5000 cells

Data independent acquisition (DIA) has emerged as a powerful proteomic technique with exceptional reproducibility and throughput, and has been widely applied to clinical sample analysis. DIA approaches normally rely on project-specific spectral libraries generated by data dependent acquisition (DDA), requiring extensive off-line fractionation and large amounts of input material. In this study, we aimed to explore the utility of DIA for the analysis of samples with limited quantities. We employed three software tools (DIA-NN, Spectronaut, and EncyclopeDIA) for data analysis and generated three types of libraries, including an experiment-specific library built by DDA analysis of off-line fractions, a FASTA sequence database, and a library generated by gas-phase fractionation (GPF), resulting in eight analysis pipelines. Then we systematically compared the performance of the eight strategies by analyzing the DIA data from HEK293T cell tryptic peptides with sample loads of 500 ng, 100 ng, 20 ng, and 4 ng. The results showed that DIA-NN with GPF-based libraries outperformed the others in protein identification and retention time calibration. Next, we further evaluated the optimized workflow by analyzing the proteome alteration in 5000 peripheral blood mononuclear cells (PBMCs) induced by lipopolysaccharide (LPS) stimulation. As a result, 3179 protein groups were quantified, and functional analysis revealed activation of multiple signaling pathways, e. g., endocytosis, NF-kappa B signaling, and T cell receptor signaling. The results showed the practicability of using DIA for scarce samples, and the established workflow of PBMC analysis could be easily adapted for biomarker discovery, immune status evaluation, and drug response monitoring, especially for diseases involved with dysfunction of the immune system.