Sensitive direct detection of cancer antigens enabled by user-defined peptide libraries

Data dependent mass spectrometry (MS) is routinely used to identify HLA-bound peptides but it can have limitations in sensitivity and reproducibility. We introduce Pepyrus, a method that uses E. coli to generate large-scale user-defined peptide libraries that can be utilized to improve the confidence in identification of HLA-bound peptides, including lowly abundant neoantigens. Using a Pepyrus peptide library paired with an HLA-specific data independent acquisition (DIA) MS strategy, we recovered >75% of the expected sequences per single injection for libraries of >10,000 peptides. Pepyrus peptide libraries also enabled the identification of 0.1 fmol spiked-in peptides in a complex background. Application of Pepyrus to create personalized peptide libraries facilitated the identification of clinically relevant HLA antigens from melanoma and renal cell carcinoma patient derived cell lines, several of which were previously undetected. Pepyrus customization enables rapid creation of patient- or disease-specific peptide libraries facilitating the confident identification of rare HLA peptides from immunopeptidomics data and the generation of large training datasets to improve spectrum, retention time, and IM prediction tools. Overall design: Replicates of pepyrus libraries encoding 10,000 peptides were amplified upstream of the peptide coding sequence and subjected to next generation sequencing.