Specificity of Aberrant Exons in a Saturated Background of Differentially Expressed Exons

Determining the maximal specificity of aberrant exons from multifactorial disease tissues in a whole body background could provide insights into the exon origin as well as potential targets of diagnosis or therapy, but it remains a challenging task. For this purpose, we obtained a saturated list of differentially expressed exons (DEXs) from our independent reference human exome extracted from 56 normal tissues by DEXSeq. We found that the DEXs comprised the majority of the reference exome and that 99.4% of cancer-specific exons relative to paired normal tissues fell within the DEX list, consistent with the ectopic expression of the majority of aberrant exons. We then screened over seven thousand pathogenic single nucleotide variants (SNVs) in the TCGA and COSMIC databases by SpliceAI. The analysis identified over three hundreds of highly confident, mostly somatic SNV-specific splicing events and/or novel exonic fragments in five types of adenocarcinomas. Interestingly, these events are all associated with cis -acting mutations in tumor suppressor genes, particularly in TP53 with an antigenic novel peptide. This revelation of the DEX exome dominance and non-specificity of nearly all aberrant exons not only reshapes our understanding of the normal human exome by highlighting its mainly variable over constitutive exons, but also illuminates the path through cis -acting somatic and pathogenic SNVs to identify genuine cancer-specific exonic fragments that are absent in any normal tissues, beyond neo-exon junctions. This path holds promise for identifying novel peptide fragments for antigens in holistic cancer treatments including vaccination and immunotherapies.