Immunotherapy has shown great therapeutic potential for cancers with high tumor mutational burden (TMB), but much less promise for cancers with low TMB. One primary approach for adoptive lymphocyte transfer-based immunotherapy is to target the somatic mutated peptide neoantigens and cancer testis (CT) antigens recognized by cytotoxic T cells. Here, we employed mass spectrometry (MS)-based proteogenomic large-scale profiling to identify potential immunogenic human leukocyte antigen (HLA) Class ǀ- associated peptides in both melanoma, a “hot tumor”, and EGFR mutant lung adenocarcinoma, a “cold tumor”. We uncovered 19 common driver oncogene-derived peptides and more than 1000 post-translationally modified peptides (PTM) representing 58 different PTMs. We constructed a CT antigen database with 286 antigens by compiling reputed CT antigen resources and “in-house” genomic data and used this to identify 45 CT antigen-derived peptides from the identified HLA peptidome. Using integrated next generation sequencing data, we discovered 12 neopeptides in EGFR mutant lung cancer cell lines. Finally, we report a novel approach for non-canonical peptide discovery, whereby we leveraged a deep learning-based de novo search and a high confidence annotated long noncoding RNA (LncRNA) database to identify 44 lncRNA-derived peptides. Findings of this study, for the first time, provide evidence for a large pool of actionable cancer antigen-derived peptides for use in mutant EGFR lung cancer immunotherapy.