Serum-antibody Profiling of H3N2-infected Ferrets Using a Combinatorial Phage-display Random Peptide Library.

The repertoire of antibodies in serum, known as the "IgOme", is highly diverse and unique to each individual as it reflects the cumulative history of personal encounters with pathogens. Consequently, profiling this repertoire may serve as a diagnostic tool for human viral infections. To explore this potential, we previously developed a computational pipeline called Motifier. The pipeline relies on random peptide sequences affinity-selected by monoclonal antibodies, demonstrating that the specifically amplified peptides can act as markers for the antibodies they bind. In this study, we evaluated whether Motifier is applicable to highly complex biological samples such as serum, which contain vast collections of antibodies, and whether biological conditions can be identified through serum-profiling. As a model system, we analyzed sera from ferrets infected with H3N2 influenza A strains. Our analyses revealed two principal findings: (i) each ferret displayed a strong and distinct antibody signature, highlighting the dominance of baseline "personal" repertoires; and (ii) peptide-motif markers associated with infection could be identified. Using these infection-related markers, we built a Random Forest classifier, which demonstrated that the markers not only characterized the biological condition but also enabled accurate prediction of unseen samples.