Cancer etiology remains an intensively studied subject, as few therapies exist to effectively treat cancer. Treating cancer is difficult because cancers arising from different tissues or individuals may be driven through entirely unique, complex mechanisms. In general, cancer research aims to address questions regarding how mutations change protein function, which mutations cause the cancer phenotype, which DNA lesions become mutations, and why the lesions are not properly repaired.

The research presented here focuses on addressing which lesions become mutations and why they are not repaired. We examine both alkylation damage and mutations, as well as UV-induced lesions and mutations, in yeast and human systems, respectively. The alkylated lesion data was produced using personally developed techniques to map lesions and repair genome-wide, and has the flexibility to be adapted for other lesions. Using these data, and other publicly available repair data, we showed that chromatin affects mutation density on a local scale, as well as the global scale. We observe increased mutation density and reduced repair at nucleosome dyads, which are least accessible to repair, causing a mutational curvature across nucleosomes. In our yeast system we also revealed that the nucleotide excision repair pathway appears to substitute for base excision repair when it is compromised, which may explain mutational asymmetries observed in certain cancers. Our human data suggest that UV lesion formation likely drives striking ~10 bp oscillatory mutational patterns in melanoma tumors across strongly positioned nucleosomes. This corresponds to the rotational setting of DNA and indicates that even unassuming genomic features can affect mutagenesis in cancers. However, repair still plays a role in creating the observed patterns. Lastly, we developed a bioinformatic pipeline to aid in examining repair pathway usage for double strand breaks. The pipeline corrects classification errors in standard alignment tools by performing local realignment, and reports information such as microhomologies that may mediate deletion events. This pipeline has been used by other labs to elucidate the mechanisms likely involved in certain cancers, such as some breast cancers.