Human proteome-wide characterization of highly mutated regions presented by HLA class I

Abstract

Tumor evolution is influenced by the interaction between cancer cells and the patient’s adaptive immune system, which shapes the tumor growth and the landscape of observed somatic variants. However, the overall impact of the immune system on the genetic variants observed in human tumors remains unclear. This study aimed to investigate regions of human proteome, here referred to as hotspot regions, that are highly presented by HLA class I molecules. The analysis was based on the presentation estimates obtained for all potential neoepitopes generated from 74 million missense variants and evaluating 246 human HLA class I alleles. Our first objective was to establish a clear definition of hotspot regions to later detect them within protein coding genes. We identified 13.974 hotspot regions, their coordinates, lengths and distribution within the human proteome. To achieve a better comprehension of these regions, we analyzed their biochemical properties and sequence composition, which revealed patterns of amino acid enrichment and depletion, specific sequence motifs, and their associations with features such as intrinsically disordered regions.