166P Development of patient-derived tumor organoids from cancers of unknown primary carcinomas to guide clinical decision making (OCAPI)

predicted antigen presentation, and "promiscuity" scores quantified peptide feature frequency.We evaluated our model's clinical relevance by examining the overall survival of 311 renal cancer patients using an SVM model based on contig-overlapping mutation data.Results: Promiscuity in T-cell antigen landscapes refers to the flexibility of MHC molecules to bind various peptides and present them by multiple alleles.Understanding this is crucial for prioritizing cancer vaccines and assessing neoantigens.The CARMEN database offers comparative antigen binding profiles across peptides, MHC alleles, populations, and genomic regions.Simplified motif landscapes from MHC Class I peptides show that over one-third of peptides are shared between two individuals.Recognizing individual and population-level promiscuity aids in selecting neoantigens for cancer patients, with CARMEN providing accessible mass-spectrometry data for the genomics community.Conclusions: Our work broadens "promiscuity" to include haplotypes, peptides, and genomic regions.We show that these regions predict immunotherapy responses and immune visibility.Validating these findings with multi-omic approaches enhances patient-specific vaccines, revolutionizing immunopeptide identification and vaccine design by integrating multi-omic data for wider coverage.