Proteome-wide discovery of novel diagnostic antigens for coccidioidomycosis

Abstract

Coccidioidomycosis, commonly called Valley Fever (VF), caused by Coccidioides spp., is an understudied fungal disease that accounts for a substantial fraction of pneumonias in endemic areas. Diagnosis remains challenging because the mainstay of diagnosis is serology, and current assays have modest sensitivity in early infection. Discovery of new antigens could improve serologic detection early in disease. Here, we use PepSeq, a DNA-programmable serology platform, to represent the complete proteome of Coccidioides within a library of 244,000 DNA-barcoded peptides. By profiling antibody reactivity at the protein and epitope levels in VF patients, naturally infected pigtail macaques, and healthy controls, we identify numerous previously-unrecognized antigens and epitopes, including some shared between humans and macaque responses. When translated into a pooled peptide ELISA, we show that a subset of identified antigens can discriminate VF cases from controls in an independent cohort with high accuracy (AUC = 0.91). These findings establish high-dimensional peptide serology as a scalable approach for fungal antigen discovery and identify new targets for improved VF diagnosis.