An Investigation of DNA Regions in Saccharomyces cerevisiae Targeted by Cisplatin or UV for Crosslinking and Creating a Draft Genome for W303

Abstract

Both cisplatin and ultraviolet (UV) light are mutagens that humans are commonly exposed to, UV in the form of natural light and cisplatin as a frequently used chemotherapy agent. Because of this it is important to understand the regions of the genome that are targeted by these mutagens. The objective of this experiment and analysis was to identify regions of DNA targeted by either Cisplatin or UV, both at the genomic level and at the basepair level. This experimental approach utilizes full genome sequencing, allowing for analysis beyond the in vitro experiments that have already been done with these mutagens. Diploid and haploid strains of Saccharomyces cerevisiae were treated with either UV or Cisplatin over 80 generations, after which their entire genomes were sequenced in order to identify DNA damage and rearrangements. From 917 mutations in UV-treated strains, six motifs around UV-damaged sites were generated. These motifs, which were largely made up of pyrimidine dimers, indicate the target sequences of UV. From 104 mutations in Cisplatin-treated strains, two motifs were generated. These motifs consisted primarily of adenine-guanine or guanine-guanine basepairs; however the results are less definitive than the UV motifs. Motifs were generated using MEME. Mutations from both Cisplatin and UV were also used to look at the fold enrichment of mutations against all features contained in the general feature format file for S288C. In addition, by tracking losses of heterozygosity using a Hidden Markov Model (HMM) in diploids we were able to identify the approximate the location and the number of interstrand crosslinks caused by both Cisplatin and UV. We found significantly more losses of heterozygosity (LOH) in the rad14-/- strains of yeast both for Cisplatin and UV treatment, indicating that nucleotide excision repair (NER) played a role in repairing crosslinks. As part of the genetic analysis of loss of heterozygosity and mutation accumulation we needed to improve upon the W303 draft genome. W303 shares 85 percent of its genome with S288C so we created an iterative assembler that takes advantage of the similarities between W303 and S288C to create a better reference genome for W303 with only short reads while simultaneously updating the general feature format file. We used Bowtie2 for alignments, SOAP for assembly, and Freebayes for variant detection.