ABSTRACT Programming the immune system to detect neoantigens and destroy tumors is critical for effective immunotherapy. Until now, bioinformatic prediction of neoepitopes on tumors from Next Generation Sequencing (NGS) information has been used alone or in conjunction with immunological assays to indirectly infer neoepitope identification. Unfortunately, only a small fraction of predicted epitopes are surface-displayed as HLA-bound peptides (pMHC), a process required for cytolytic T lymphocyte (CTL) targeting. Moreover, immunologic assays suffer from both high false positive and false negative rates, confounding correct identification. Conventional mass spectrometry (MS) approaches to interrogate the pMHC, referred to as the cell's immune peptidome, suffer from poor HLA recovery, requirement for multiple sample runs to achieve adequate peptide coverage and necessitate large numbers of tumor cells, all features impractical for routine clinical use. Our Academic-Industrial Partnership (AIP) advances the creation of a commercial pipeline to deliver personalized tumor neoantigen identification, integrating NGS-based genomics and transcriptomics, bioinformatics, chemical peptidomics and a novel, ultrasensitive form of MS. Our interdisciplinary/multi-institutional strategic alliance combines basic research at Dana Farber Cancer Institute with industrial expertise at Curacloud Corporation and JPT Peptide Technologies. We propose deployment of an attomole (10-18) Poisson detection liquid chromatography-data independent acquisition (LC-DIA) MS method for antigen discovery to electronically record and capture the entire immune peptidome comprising both numerous self-peptides and sparse neoantigens in a single run from small numbers of tumor cells (106) retrieved by clinical needle biopsy. This approach changes the aforementioned MS calculus and permits neoantigen search at any point following data collection using existing commercially marketed MS instrumentation. In Aim 1 neoepitope candidates shall be chemically synthesized in high throughput pools of up to 6,000 peptides per nanoscale run by JPT for MS fragmentation analysis and elution mapping reference standards for definitive neoantigen identification using LC-DIAMS on individual tumor samples based on DFCI technology, optimizing each step. In Aim 2 we shall use NGS data from tumor cells in conjunction with bioinformatics at Curacloud to predict neoepitopes arising from coding and non-coding regions capable of interacting with each HLA-A, -B and/or -C allele of a patient. Machine learning-based neoepitope ranking algorithms incorporating MS data and other results shall be developed for candidate prioritization. An end user service shall be established involving all aforementioned integrative technologies. From initial tumor biopsy to identification of neoepitopes, a time scale of approximately one month is anticipated. This generic neoepitope precision identification pipeline is applicable to multip...