A Viral Mutation Profiling and Discovery Strategy for Sensitive Multiplex Detection of Viruses and Variants in Saliva by Proteomics

Abstract Detecting multiple viruses in biofluids such as saliva remains challenging because viral proteins are of low abundance, embedded in a complex matrix, and subject to rapid sequence variation. Mass spectrometry-based proteomics complements nucleic acid–based assays by directly resolving protein-level sequence variation, but standard database cross-correlation searches lose sensitivity as viral variant sequences increase the database size. Here, we report a sequence-filtering strategy integrated with single amino acid variant (SAAV) detection to enable sensitive mutation variant detection while minimizing the loss of search sensitivity. In silico analysis of SARS-CoV-2 and influenza sequences reveals bimodal prevalence distributions of the mutated peptides. By using a three-month time window and removing low-prevalence peptides, the peptide database size is reduced by 82.1% for SARS-CoV-2 and 94.6% for influenza at a 10% prevalence threshold with only a <0.8% loss in search sensitivity while retaining 97.8% and 98.4% of viral peptides in the subsequent month. A mutation discovery module further mitigates loss of database coverage and improves variant detection in peptide-spiked saliva experiments compared with de novo sequencing, achieving SAAV peptide-spectrum match recovery rates of 73.6% for the SARS-CoV-2 spike and 88.4% for H1N1 hemagglutinin. This study establishes a scalable proteomics framework for the sensitive detection of known and emerging viral variants in complex biofluid matrices.