Deep Data-independent Acquisition-based Plasma Proteomic Profiling Unveils Distinct Molecular Features in Dengue Fever with Neutropenia

Methods

Plasma samples were collected from dengue fever (DF) patients, DF patients with neutropenia (DFN), and healthy controls (HC). A comprehensive plasma proteomic profiling was conducted using a deep data-independent acquisition (DIA) workflow combined with LC-MS/MS analysis. Differential protein expression analysis and bioinformatic approaches, including functional enrichment analyses and machine learning, were utilized to elucidate key cellular pathways and identify promising biomarkers.Results: DFN patients exhibited significant dual hematological alterations, with notable changes in both platelet and neutrophil counts, reflecting a complex disturbance in hematological homeostasis during dengue progression.

DIA analysis quantified 2,475 proteins, revealing widespread proteomic alterations among the DF, DFN, and HC subjects. Differential analysis highlighted significant fluctuations in proteins related to cytoskeletal organization, metabolic regulation, and intracellular signaling. Enrichment analyses implicated pathways such as focal adhesion, platelet activation, and PI3K-Akt signaling.

Machine learning methods further identified a panel of four biomarkers—CNST, DSTN, DUSP3, and PDIA5—with high predictive accuracy for dengue diagnosis and subgroup differentiation.Conclusion: This study advances our understanding of dengue’s plasma proteomic landscape and underscores the synergistic potential of DIA-based proteomics and machine learning in unveiling host-response mechanisms, thereby informing early diagnosis and targeted therapeutic strategies.