SOMATIC MOSAICISM FOR SINGLE NUCLEOTIDE VARIANTS IN HUMAN TISSUES OF APPARENTLY HEALTHY INDIVIDUALS: A MULTI-TISSUE ANALYSIS OF SOMATIC & CANCEROUS MUTATIONS

Abstract

Cancer is a disease which can be caused due to hereditary traits, exposure to environmental mutagens, and/or mutations arising due to DNA replication errors. Recent studies suggest that cancerous somatic mutations occur at a much higher frequency in apparently healthy individuals than previously appreciated, making this a promising yet understudied area of research in cancer genomics as most studies look only at tumor samples. In this study, we evaluated the relationship between the number of general and cancerous somatic mutations and the following variables: age of donor, sex of donor, organ/tissue the sample came from, and the self-renewal capacity of that organ/tissue. Specifically, we used whole exome sequencing methods to call somatic single nucleotide variants (SNVs) in the protein coding regions of the genomes from tissue samples of 14 apparently healthy donors (ages ranging from 20-69; number of tissue samples for each donor ranging from 10-29). We identified patterns of mutation prevalence across the previously mentioned variables, looking specifically at somatic and cancerous SNV counts and allelic frequencies. We also ran multiple linear regression models to estimate the effect of these predictor variables on the number of somatic and cancerous SNVs. We found, in total, 3,032 somatic SNVs in our 266 total tissue samples, as well as 162 cancerous somatic SNVs (meaning they overlapped in genomic position with known pathogenic cancerous mutations from the COSMIC database). The data did not find any statistically significant effect of age, sex, or organ on the number of somatic and cancerous SNVs, but did find a statistically significant effect of self-renewal capacity of the organ/tissue sample on the number of both somatic and cancerous SNVs. Our plots indicate differences in mutational patterns (general somatic and cancerous SNVs) across organ/tissue of sample origin as well as across donor. Together, these results challenge long held assumptions in cancer genomics, such as the appropriateness of the current classification method of cancer (based on organ/tissue of origin), as well as the assumption that age leads to accumulation of mutations continuously over time. Moreover, they underscore the importance of considering effects of tissue level processes (such as self-renewal) on the patterns of mutations and the importance of investigating multi-tissue patterns of mutations in healthy individuals as a means to improve early detection methods and to better analyze the risk and prognosis of cancer.