Title: Chromatin marks govern mutation landscape of cancer at early stage

7 * Co-corresponding authors 8 9 10 Accumulation of somatic mutations over time leads to tissue abnormalities, such as cancer. 11 Somatic mutation rates vary across the genome in a cell-type specific manner, depending on 12 the types of mutation processes 1-7. Although recent studies have identified several 13 determinants relevant to the establishment of the cancer mutation landscape 8-13 , these studies 14 have yet to propose the major time point at which these factors come into play during cancer 15 progression. Here, we analyzed whole genome sequencing data from two different types of 16 precancerous tissues, monoclonal B-cell lymphocytosis and Barrett's esophagus, and their 17 matching cancer types along with 423 epigenetic features from normal tissues to determine 18 the critical time point when chromatin features contribute to the formation of the somatic 19 mutation landscape. Our analyses revealed that a subset of cell-of-origin associated 20 chromatin features can explain more than 80% of the regional mutation variance for both 21 types of precancerous tissues, comparable to the variance explained level for the genomes of 22 matching cancer types. In particular, major significant chromatin features explaining the 23 mutation landscape of Barrett's esophagus and esophageal adenocarcinoma were derived 24 from stomach tissues, indicating that mutation landscape establishment occurs mostly after 25 environment-mediated epigenetic changes during gastric metaplasia. Analyses of the genome 26 of esophageal squamous cell carcinoma tissues demonstrated that the proposed time point for 27 mutation landscape establishment of Barrett's esophagus and esophageal adenocarcinoma 28 were specific to the occurrence of cell-type shift. Thus, our data suggest that the major time 29 point for the mutation landscape establishment dictated by chromatin features is early in the 30 process of cancer progression, and epigenetic changes due to environmental conditions at 31 early stages can dramatically impact the somatic mutation landscape of cancer. 32. CC-BY-NC-ND 4.0 International license not peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was. Recent advances in cancer genomics have so far revealed numerous somatic mutation 33 landscapes for various cancer types, leading to a number of key findings. Identification of 34 new driver gene mutations, deciphering clonal evolution structure and profiling tumor 35 heterogeneity within and among different patients through examination of mutations, mainly 36 at the gene level 1-7 , have successfully addressed the genes contributing to cancer progression 37 and identified novel therapeutic …