Pepsickle rapidly and accurately predicts proteasomal cleavage sites for improved neoantigen identification

Abstract

Proteasomal cleavage is a key component in protein turnover, as well as antigen presentation and subsequent immune response. Herein we present pepsickle, an open-source tool for proteasomal cleavage prediction with better in vivo prediction performance (AUC) and computational speed than current models available in the field, and with the ability to predict sites based on both constitutive and immunoproteasome profiles. Post-hoc filtering of predicted patient neoepitopes using pepsickle significantly enriches for immune-responsive epitopes and may represent a significant opportunity to improve current epitope prediction and vaccine development pipelines.