AI-Guided Proteomic Biomarker Panel for Differentiating Bacterial and Viral Infections in Acute Febrile Illness

This study aims to identify and validate proteomics-based diagnostic biomarkers for differentiating bacterial from viral acute febrile illnesses. It is conducted on the same patient cohort previously used in a registered transcriptomic biomarker study (ClinicalTrials.gov Identifier: NCT065529754). Although both studies share the overarching clinical objective of improving infection triage, they are scientifically independent in terms of methodology, biomarker modality, and analytical pipeline. Transcriptomic study (NCT065529754): Focused on host-response gene expression signatures derived from RNA sequencing data. Proteomic study (this registration): Focuses on host-response protein biomarkers identified through high-resolution mass spectrometry (DIA/PRM/SRM) and validated by immunoassays (ELISA). The proteomics study is designed to discover a minimal and biologically distinct set of protein markers that can be readily translated into clinical diagnostics. This study evaluates a minimal set of circulating proteins-ICAM1, CFHR5, and GRN-discovered through AI-assisted proteomics and validated by ELISA, as a rapid diagnostic tool for bacterial vs viral infection. Patients with acute fever are prospectively enrolled into three cohorts: discovery, internal validation, and external validation. This independent registration reflects the proteomics-specific objectives, methodology, and outcomes, while also acknowledging its linkage to the transcriptomics study.